Holometabolous insects (Holometabola) Brian M. Wiegmann*, Jung-wook Kim, (legs, wings, antennae, genitalia) then develop from spe- Michelle D. Trautwein cialized internal regions of subcuticular epidermal cells Department of Entomology, North Carolina State University, Raleigh, called imaginal discs (4). 7 e larval cuticle is reduced or NC 27695-7613, USA entirely lost and an adult cuticle is newly formed. 7 e *To whom correspondence should be addressed internal development of the wings is denoted in the other ([email protected]) common name of the group, the Endopterygota. Despite its huge diversity, there are relatively few holometabolan Abstract lineages that contain exceptionally large numbers of spe- cies. Developmental specialization clearly played a major The Holometabola includes 11 orders that represent the role in the expansion of holometabolan life histories, but vast majority of insect diversity (~850,000 species). Recent the hyperdiversity of major clades of beetles, P ies, moths, molecular and morphological treatments support holom- and wasps are most oJ en attributed to independent, lin- etabolan monophyly, confi rm the monophyly of the major eage-speciA c radiations enabled by unique combinations orders, and provide new evidence to place the orders in a of trophic, life history, morphological adaptations, and phylogeny. Estimates of divergence time based on molecu- the expansion of terrestrial plant communities ( 2, 5–9). lar evidence suggest an origin of the Holometabola within For a more comprehensive perspective on insect diver- the Carboniferous, 359–299 million years ago, Ma, but sity, fossil history, and evolution, see Grimaldi and Engel defi nitive fossils fi rst appear in the Permian, 299–280 Ma. (2). Here, we review evidence on the phylogeny and The molecular timetree reveals striking parallel radiations divergence times of holometabolous insects. of insect lineages throughout the Mesozoic (251–66 Ma). Phylogenetic classiA cations of Holometabola based on morphological features divide the group into two major 7 e insect clade Holometabola (~850,000 species) subclades, the Neuropteroidea, which includes Cole- includes 11 living orders that together comprise the vast optera + the Neuropterida (Neuroptera, Megaloptera, majorit y of a ll insect diversit y and therefore a lso represent Rhaphidioptera), and the Mecopterida (= Panorpida), a signiA cant fraction (>60%) of all terrestrial animals (1). Holometabola includes the four largest orders of insects: Coleoptera (beetles, Fig. 1), Hymenoptera (bees, ants, and wasps), Diptera (true P ies), and Lepidoptera (moths and butterP ies), as well as the Neuroptera (lacewings), Megaloptera and Raphidioptera (dobson- P ies and alderP ies), Trichoptera (caddisP ies), Mecoptera (scorpionP ies), Siphonaptera (P eas), and Strepsiptera (twisted-wing insects). 7 e name of the group reP ects their deA ning characteristic—they undergo complete metamorphosis. 7 eir life history is divided into discrete developmental stages, including a distinct larval (feed- ing) and pupal (quiescent) stage. 7 e major developmen- tal, morphological, and behavioral modiA cations that led to the holometabolous larva are thought to have arisen through extension of the prenymphal stage of hemi- metabolous insects (2, 3). Metamorphosis from larval to Fig. 1 A holometabolous beetle larva (Coleoptera: adult morphology occurs in the pupal stage where the Chrysomelidae, Zygogramma sp.) from Arizona, USA. Photo larval structures are broken down and adult features credit: A. Wild. B. M. Wiegmann, J. Kim, and M. D. Trautwein. Holometabolous insects (Holometabola). Pp. 260–263 in e Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009). HHedges.indbedges.indb 226060 11/28/2009/28/2009 11:27:21:27:21 PPMM Eukaryota; Metazoa; Arthropoda; Insecta; Holometabola 261 Mecoptera 9 Siphonaptera 7 Diptera 5 Antliophora Lepidoptera 10 Mecopterida Trichoptera 3 Neuroptera Amphiesmenoptera 8 Rhaphidioptera 11 2 4 Megaloptera Coleoptera 1 6 Strepsiptera Neuropteroidea Hymenoptera Hemimetabolous outgroups C P Tr J Cretaceous Pg Ng PALEOZOIC MESOZOIC CZ 350 300 250 200 150 100 50 0 Million years ago Fig. 2 A timetree of the holometabolous insects. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), J (Jurassic), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic). including Lepidoptera, Trichoptera, Diptera, Mecop- alignment, and methods of analysis. Previous molecular tera, and Siphonaptera. Hymenoptera and Strepsiptera studies, most using rDNA, have recovered a monophy- have been placed in various positions, the former oJ en letic Neuropterida (16, 17), Neuropteroidea (16, 18, 19), placed as closest to the Mecopterida and the latter trad- Amphiesmenoptera (16, 17, 19–21), Mecopterida (20, 21) itionally placed either within or closest to Coleoptera (10, and, most provocatively, Halteria (Strepsiptera + Diptera) 11). 7 e consensus view is that most morphological fea- (17, 19). Two recent phylogenomic projects, with limited tures of the Hymenoptera and the Strepsiptera are too taxon sampling but large numbers of genes, addressed highly modiA ed to unequivocally resolve their phylo- the placement of the Hymenoptera; mitochondrial genetic positions. Additional widely accepted groupings genomes provided evidence that Hymenoptera is the are the Amphiesmenoptera (Lepidoptera + Trichoptera) closest relative of Mecopterida (22), while combined ana- (8) and the Antliophora (Diptera + Mecoptera including lysis of 185 nuclear genes strongly supports placement of Siphonaptera) (12). Morphological evidence also sup- the Hymenoptera as the earliest branching lineage, the ports a close relationship between the Mecoptera and closest relative of all other Holometabola (23). the Siphonaptera (11). 7 e most comprehensive review 7 e most current molecular study by Wiegmann of the morphological evidence for holometabolan rela- et al. (submitted) is the A rst to include both nuclear tionships is that of Kristensen (1999; 8) and was further genes (AATS, CAD, TPI, SNF, PGD, and RNA POLII) evaluated in light of emerging alternative phylogenetic and representative taxa from all holometabolan orders. hypotheses by Beutel and Pohl (12). New perspectives 7 eir A ndings support traditional morphological on speciA c character systems such as sclerites, muscle hypotheses (Neuropteroidea + Mecopterida includ- insertions, and functional features of the wing-base (13), ing Amphiesmenoptera + Antliophora) and Savard and mouthparts (14), as well as new paleontological A nd- et al.’s (23) early branching position for Hymenoptera. ings and interpretations (2, 15) continue to add to the Additionally, multiple nuclear genes provide evidence evidence on relationships. for the placement of the enigmatic Order Strepsiptera Molecular analyses of holometabolan phylogeny as the closest relative of Coleoptera. 7 ese results add have primarily relied on 18S ribosomal DNA, and the to the compounding and conP icting evidence for the results have been highly dependent on taxon sampling, placement of Strepsiptera—the most controversial issue HHedges.indbedges.indb 226161 11/28/2009/28/2009 11:27:22:27:22 PPMM 262 THE TIMETREE OF LIFE Table 1. Divergence times (Ma) and their credibility/ both dipterans and lepidopterans and found the origin confi dence intervals (CI) among holometabolous of the taxon-limited Holometabola to be 351–338 Ma insects, based on Wiegmann et al. (submitted). (38). Wiegmann et al. (submitted) estimated the diver- Timetree gence times of all holometabolan orders using a Bayesian Node Time CI phylogeny based on multiple nuclear genes, fossil cali- brations, and relaxed clock Bayesian methods using the 1355360–334program Multidivtime (39). Congruent with the A ndings 2350359–336of Gaunt and Miles (38), multiple nuclear genes placed 3300315–287the origin of the Holometabola at 355 Ma, within the 4286299–274Carboniferous, but earlier than traditional estimates. 5282300–2647 e Hymenoptera, as the earliest branching lineage in 6 274 285–270 the phylogeny, has an age of origin nearly equivalent with 7256279–234the age of the divergence of Holometabola from its closest relative (Fig. 2). 7 is date is considerably older than exist- 8255276–227 ing fossil estimates, typical of molecular estimates (39). 9 243 270–222 7 e split between the two major subclades Neuropteroidea 10 230 261–190 and Mecopterida took place in the Permian 300 Ma, with 11 213 247–134 the Amphiesmenoptera/Antliophora divergence occur- ring 284 Ma. 7 e divergence of all orders (excluding the Hymenoptera) appears to have occurred in relatively rapid succession, with dates of origin falling in the range in holometabolan phylogenetics. Beutel (12) recently 274–213 Ma, with the earliest being the Coleoptera/ reviewed arguments surrounding the “Strepsiptera prob- Strepsiptera divergence at 274 Ma and the most recent lem” (24). 7 e original A ndings of Whiting and Wheeler being the split of Rhaphidioptera and Megaloptera at (25) placed Strepsiptera as the closest relative of Diptera 213 Ma. 7 ough some estimates of ordinal-level diver- based on 18S rDNA and initiated a useful debate regard- gences do not precisely correspond with traditional ages ing empirical evidence for spurious grouping by “long- based on fossils, paleontological evidence is dramatically branch attraction” in molecular phylogenetics (26–30). expanding, and thus, better fossil calibrations coupled Further studies supporting Halteria were based on 28S